Brain natriuretic peptide (BNP 1–32) plays an important physiologic role in cardiorenal homeostasis. Recently, it has been reported that BNP 1–32 is rapidly cleaved by the ubiquitous enzyme dipeptidyl peptidase IV to BNP 3–32, which lacks the two NH2-terminal amino acids of BNP 1–32. The bioactivity of BNP 3–32 in cardiorenal regulation is unknown. We hypothesized that BNP 3–32 has reduced vasodilating and natriuretic bioactivity compared with BNP 1–32 in vivo. Synthetic human BNP 3–32 and BNP 1–32 were administered to eight anesthetized normal canines. After baseline measurements, BNP 1–32 at 30 ng·kg−1·min−1 was administered, followed by a washout, a postinfusion clearance, and a clearance with an equimolar dose of BNP 3–32. In four studies, the sequence of BNP 1–32 and BNP 3–32 infusion was reversed. Peptides were compared by analyzing the changes from the respective preinfusion clearance to the respective infusion clearance. *P < 0.05 between peptides. BNP 3–32, unlike BNP 1–32, did not decrease mean arterial pressure (0 ± 1 vs. −7 ± 2* mmHg, respectively) and did not increase renal blood flow (+12 ± 10 vs. +52 ± 10* ml/min). Effects on heart rate and cardiac output were similar. Urinary sodium excretion increased 128 ± 18 μeq/min with BNP 3–32 and 338 ± 40* μeq/min with BNP 1–32. Urine flow increased 1.1 ± 0.2 ml/min with BNP 3–32 and 2.8 ± 0.4* ml/min with BNP 1–32. Plasma BNP immunoreactivity was lower with BNP 3–32, suggesting accelerated degradation. In this study, BNP 3–32 showed reduced natriuresis and diuresis and a lack of vasodilating actions compared with BNP 1–32.
- enzymatic degradation product
- cardiorenal regulation
- BNP 3–32
brain natriuretic peptide (BNP 1–32) is a member of the natriuretic peptide family, which also includes atrial natriuretic peptide (ANP) and C-type natriuretic peptide. BNP 1–32 binds to the natriuretic peptide type A receptor (NPR-A) and via the second messenger cyclic guanosine monophosphate (cGMP) mediates a variety of actions including vasodilation, natriuresis, suppression of renin secretion, lusitropism, and inhibition of fibrosis (3, 21, 22). Under physiological conditions, the plasma concentrations of BNP are very low; however, they can increase markedly with cardiac overload, which has led to its use as a cardiac biomarker (13).
Mature BNP 1–32 is cleaved from the 108 amino acid prohormone BNP, presumably by the serine protease corin (24). Subsequently, BNP 1–32 either binds to the NPR-A or the natriuretic peptide clearance receptor or is enzymatically degraded. While ANP is a substrate for degradation by neutral endopeptidase 24.11, BNP appears to be relatively resistant (4, 9, 14). Most recently, Brandt et al. (4) reported that the ubiquitous aminopeptidase dipeptidyl peptidase IV (DPP4; CD26; EC 18.104.22.168) cleaves BNP 1–32 to produce BNP 3–32 (Fig. 1). Importantly, the specificity constant of BNP 1–32 was comparable to those reported for the DPP4 substrates glucagon-like peptide-1 and glucose-dependent insulinotropic polypeptide, for which cleavage by DPP4 in vivo has been reported (4, 10). Truncation of BNP 1–32 to BNP 3–32 is also consistent with the presence of BNP 3–32 in human plasma reported by Shimizu et al. (20) who had sought to define additional molecular forms of the BNP system in human heart failure. However, what remains unclear is the biological activity of BNP 3–32. Specifically, it is undefined in vivo if removal of two NH2-terminal amino acids from BNP 1–32 enhances, attenuates, or has no effect on biological activity in the overall control of cardiorenal function. In addition to being of physiological importance, this question has now gained important clinical significance as DPP4 activity has been reported to be increased in cardiovascular disease states and DPP4 inhibitors are in clinical development for the treatment of diabetes mellitus, as several peptides involved in glucose homeostasis are also substrates of DPP4 as mentioned above (7, 10, 11, 15).
Therefore, we designed the current study to define for the first time the cardiorenal actions of BNP 3–32 and to compare it to mature biologically active BNP 1–32. To address this aim, renal, cardiovascular, and endocrine functions were assessed in a large animal model that permits assessment of integrative physiologic function. Our major findings were that compared with mature BNP 1–32 the renal excretory actions of BNP 3–32 were significantly attenuated and the vasodilating actions absent.
MATERIALS AND METHODS
The current study was performed in eight male mongrel dogs (weight 20–28 kg) in accordance with the Animal Welfare Act and with approval of the Mayo Clinic Animal Care and Use Committee.
Dogs were maintained on a sodium-controlled diet (Hill’s I/d diet; Hill’s Pet Nutrition, Topeka, KS). On the evening before the acute experiment, they were fasted and given access to water ad libitum. On the day of the acute study, animals were anesthetized with pentobarbital and fentanyl, intubated, and mechanically ventilated with 5 l/min supplemental oxygen. A flow-directed balloon-tipped thermodilution catheter was inserted via the right external jugular vein for hemodynamic measurements. The femoral vein was cannulated for continuous infusions, and the femoral artery was cannulated for mean arterial pressure measurements and blood sampling. Pressures were recorded and analyzed digitally (Sonometrics, London, ON, Canada). Via a left lateral flank incision the ureter was cannulated for urine sampling, and the renal artery was equipped with a flow probe (Carolina Medical Electronics, King, NC). Cardiac output was measured by thermodilution (cardiac output model 9510-A computer; American Edwards Laboratories, Irvine, CA).
The study protocol started with the administration of a weight-adjusted inulin bolus. Continuous inulin and saline infusions at a rate of 1 ml/min each were started. After 60 min of equilibration, a baseline clearance was done. All clearances lasted 30 min and consisted of urine collection, blood sampling, and hemodynamic measurements. After the baseline clearance, the saline infusion was replaced with an infusion of synthetic human BNP 1–32 (Phoenix Peptide, Belmont, CA; diluted in saline) at a concentration of 30 ng·kg−1·min−1 (infusion rate 1 ml/min). After a lead-in period of 15 min, a 30-min clearance was done. Thereafter, BNP 1–32 was replaced with a saline infusion (1 ml/min), and after a washout period of 60 min, a postinfusion clearance was done. Thereafter, the saline infusion was replaced with an infusion of synthetic human BNP 3–32 (synthesized by the Peptide Synthesis Facility, Mayo Clinic Rochester, MN; diluted in saline) on an equimolar basis compared with the BNP 1–32 infusion (i.e., 28.39 ng·kg−1·min−1, infusion rate 1 ml/min). After a 15-min lead-in period, a 30-min clearance was done. In half of the studies, the sequence of BNP 1–32 and BNP 3–32 infusions was reversed to compensate for a possible carryover effect. Volume loss was replaced with saline.
Analysis of electrolytes and neurohormones.
Electrolytes were measured by flame photometry (model IL943; Instrumentation Laboratory, Lexington, MA). Inulin was measured with the anthrone method (6). Glomerular filtration rate was assessed by inulin clearance. Plasma renin activity, angiotensin II, and aldosterone were determined by commercially available radioimmunoassays as described previously (12). Immunoreactivity for human BNP was measured with an immunoradiometric assay (Shionogi, Tokyo, Japan). cGMP was measured using a competitive RIA cGMP kit (PerkinElmer, Boston, MA).
Values are expressed as mean ± SE. For each peptide, changes from preinfusion levels were analyzed with paired t-test for normally distributed data. Peptides were compared with each other by analyzing the changes from the respective preinfusion clearance to the respective infusion clearance with paired t-test for normally distributed data. BNP values were log transformed before analysis. Wilcoxon signed rank test was used for data not normally distributed, specifically urinary sodium excretion and urinary cGMP excretion. Statistical significance was accepted at P < 0.05. Analyses were performed with GraphPad Prism 3.02 (GraphPad Software, San Diego, CA).
Cardiorenal and humoral functions are reported in Table 1 and Figs. 2 and 3. Preinfusion parameters for BNP 3–32 and BNP 1–32 were similar with the exception of plasma aldosterone, which was lower before BNP 3–32 infusion (15.0 ± 2.6 vs. 18.6 ± 3.3 ng/dl, BNP 3–32 vs. BNP 1–32; P = 0.04). No reliable cardiac output measurements could be obtained in two studies, one with BNP 1–32 and the other with BNP 3–32 as the first peptide infusion. Thus, values for cardiac output, and consequently systemic and pulmonary vascular resistances, were taken from six studies.
Mean arterial pressure (MAP) decreased significantly with BNP 1–32 (−7.3 ± 1.8 mmHg) but not with BNP 3–32 (−0.2. ± 1.2 mmHg), and this was significant between peptides (P = 0.016; Fig. 2A). Similarly, systemic vascular resistance decreased significantly with BNP 1–32 but not with BNP 3–32, and this tended to be significant between peptides (P = 0.07). Heart rate increased with BNP 3–32 and tended to increase with BNP 1–32 (P = 0.052) with no difference between peptides. Cardiac output and right atrial pressure remained unchanged with both peptides. Pulmonary artery pressure and pulmonary capillary wedge pressure decreased with BNP 1–32 but not with BNP 3–32 (P = 0.35 and P = 0.14, respectively) with no differences between peptides. In contrast, renal blood flow increased (+23 ± 2%) and renal vascular resistance decreased (−22 ± 1%) with BNP 1–32, whereas they remained unchanged with BNP 3–32 (+5 ± 2% and −3 ± 2%, respectively), and this was significantly different between peptides (Fig. 2B).
Both urine flow (Fig. 2C) and urinary sodium excretion (Fig. 2D) increased significantly compared with preinfusion levels but less so with BNP 3–32 compared with BNP 1–32 (P = 0.004 and P = 0.008 between peptides, respectively). The same was true for urinary potassium excretion and urinary cGMP excretion (P = 0.008; Fig. 3A). BNP 1–32 increased glomerular filtration rate, while BNP 3–32 tended to do so (P = 0.09) with no difference between peptides.
BNP immunoreactivity was not measured in two studies, one with BNP 1–32 and one with BNP 3–32 as the first peptide infusion, because the assay was not available. Immunoreactivity for human BNP was undetectable at baseline in five of the six studies for which data were available. BNP 3–32 infusion led to a significantly lesser increase in human BNP immunoreactivity than did BNP 1–32 infusion (33 ± 7% of the respective increase with BNP 1–32, P = 0.001; Fig. 3B). Similarly, BNP 3–32 led to a significantly lesser increase in plasma cGMP compared with BNP 1–32 (31 ± 7% of respective increase with BNP 1–32, P = 0.001; Fig. 3C). There were no differences between peptides with regard to changes in ANP, plasma renin activity, angiotensin II, aldosterone, sodium, and potassium. BNP 3–32 was associated with an increase in ANP and decreases in angiotensin II and plasma potassium, while BNP 1–32 decreased plasma renin activity, angiotensin II, and aldosterone. Hematocrit increased significantly with BNP 1–32 but remained unchanged with BNP 3–32, and this tended to be significant between peptides (P = 0.07).
This study reports for the first time that BNP 3–32, which is produced by enzymatic degradation of BNP 1–32 by the ubiquitous aminopeptidase DPP4, has diuretic, natriuretic, and cGMP-generating properties that are reduced in vivo in healthy anesthetized canines compared with equimolar BNP 1–32. In addition, BNP 3–32 lacked the vasodilating actions of BNP 1–32.
BNP 1–32 is known to have vasodilating, natriuretic, and renin suppressing actions, which were evident in our study. Importantly, BNP 3–32 had a significantly reduced ability to increase plasma cGMP, urine flow, urinary sodium excretion, and urinary cGMP excretion compared with BNP 1–32. Unlike BNP 1–32, it did not decrease mean arterial pressure, systemic or renal vascular resistance, and it did not increase renal blood flow. BNP 3–32 slightly increased plasma ANP, which was not statistically different from BNP 1–32, which also showed a trend for increasing ANP compared with preinfusion levels (P = 0.13). An increase in plasma ANP with BNP infusion has been reported earlier (5), and a possible explanation for this increase could be competition at the NPR-A receptor or at clearance mechanisms. Also consistent with previous reports is the increase in hematocrit with BNP 1–32 (16). This could be due to the diuretic effect of BNP 1–32 and to an increase in vascular permeability associated with NPR-A activation (18).
Several mechanisms could account for the reduced biological actions of BNP 3–32 compared with BNP 1–32. First, it is possible that removal of two amino acids from the NH2 terminus resulted in altered ligand-receptor interactions. This would not be surprising as previous studies with ANP have reported that deletion of amino acids on either the COOH terminus or NH2 terminus can affect ANP bioactivity (19, 23). However, we recently reported that BNP 1–32 and equimolar BNP 3–32 stimulated cGMP generation to a similar degree in cultured human cardiac fibroblasts (8), which argues against reduced receptor affinity of BNP 3–32 as an explanation for the findings in the current study. Another possible mechanism could be that removal of the two NH2-terminal amino acids of BNP 1–32 results in a molecule that is highly susceptible to further degradation or clearance or both. This is indirectly supported by our observations regarding plasma BNP-immunoreactivity and cGMP during infusion of equimolar concentrations of BNP 3–32 and BNP 1–32. Specifically, the increase in BNP immunoreactivity for BNP 3–32 was only about 33% of that observed for BNP 1–32; this was paralleled by a cGMP increase with BNP 3–32 that was only 31% of that induced by BNP 1–32. Of note, since the Shionogi assay employed in this study uses antibodies directed against epitopes on the ring structure and the COOH terminus of BNP 1–32, it can be expected, and indeed it has been reported that BNP detection with this assay is not affected by changes that are restricted to the NH2 terminus (17, 20). Thus, the assay should detect BNP 1–32 and BNP 3–32 with similar affinity. One could then speculate that enzymatic degradation of BNP 1–32 is a stepwise process, with every cleavage rendering the remaining peptide subject to one or more different peptidases with appropriate amino acid sequence-specific cleavage sites. As the BNP immunoreactivity during the BNP 3–32 infusion was only ∼33% of that seen with BNP 1–32 with a commensurate reduction in plasma cGMP generation, we speculate that cleavage of the two NH2-terminal amino acids accelerates further enzymatic degradation and thus inactivation.
The current findings have clinical relevance for the pathophysiology and therapeutics of heart failure in which BNP concentrations are high in the plasma and in which BNP 1–32 has been used as a therapeutic agent. It has been reported that DPP4 activity is increased in plasma of hypertensive patients with increased pulmonary artery pressure (15) and in atrial tissue samples of patients with chronic persistent atrial fibrillation undergoing open heart surgery (11). If DPP4 activity is increased in cardiovascular disease states, then the availability of the active mature BNP 1–32 may be reduced and its cardiorenal protective actions attenuated as well. From a treatment perspective, DPP4 inhibitors are being developed and have been extensively investigated as a potential treatment in diabetes mellitus, as the incretins glucagon-like-peptide-1 (GLP-1), GLP-2, and glucose-dependent insulinotropic peptide are some of the many substrates of DPP4 (1, 7). Of note, DPP4 inhibition did not cause hypoglycemia in healthy male volunteers (2). Thus, the possible therapeutic use of DPP4 inhibitors in heart failure may be an area warranting further research. If indeed DPP4 inhibitors were able to increase the bioavailability of BNP 1–32, it would be important to investigate whether chronic endogenous BNP augmentation could delay the progression of heart failure in less-advanced stages and whether a DPP4 inhibitor could increase the efficacy of exogenous BNP 1–32 in decompensated heart failure. Importantly, as other enzymes as well as the natriuretic peptide C receptor are involved in the clearance of BNP 1–32, studies with actual DPP4 inhibition will be required to assess the impact on BNP levels. Furthermore, given potentially relevant species differences, the findings of the current study, in which synthetic human BNP 1–32 and BNP 3–32 were given to canines, need to be confirmed in humans.
In summary, in this study BNP 3–32, the product of BNP 1–32 cleaved by DPP4, has reduced renal actions compared with BNP 1–32 and lacks vasodilating properties. These findings provide new insights into the integrated cardiorenal physiology of the BNP system with possible therapeutic implications.
This research was supported by National Heart, Lung, and Blood Institute Grants HL-36634 and POI-HL-07111 and by the Mayo Foundation.
We are very grateful to Denise M. Heublein, Sharon M. Sandberg, and Lynn K. Harstad for technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society